Online LIDAR-to-ground alignment

Abstract

A LIDAR-to-vehicle alignment system includes a sensor data collection module configured to collect points of data provided based on outputs of one or more LIDAR sensors. An alignment module is configured to identify lane markings based on the points of data, determine a lane marking direction based on the identified lane markings, calculate a yaw of a LIDAR coordinate system relative to a vehicle coordinate system based on the determined lane marking direction, and identify a ground plane based on the points of data. The alignment module is also configured to calculate a roll and pitch of the LIDAR coordinate system relative to the vehicle coordinate system based on the identified ground plane, and update a transformation matrix based on the calculated yaw, roll, and pitch of the LIDAR coordinate system.

What is claimed is: 
     
         1 . A LIDAR-to-vehicle alignment system, comprising:
 a sensor data collection module configured to collect points of data provided based on outputs of one or more LIDAR sensors; and   an alignment module configured to
 identify lane markings based on the points of data, 
 determine a lane marking direction based on the identified lane markings, 
 calculate a yaw of a LIDAR coordinate system relative to a vehicle coordinate system based on the determined lane marking direction, 
 identify a ground plane based on the points of data, 
 calculate a roll and pitch of the LIDAR coordinate system relative to the vehicle coordinate system based on the identified ground plane, 
 update a transformation matrix based on the calculated yaw, roll, and pitch of the LIDAR coordinate system, and 
 not update the transformation matrix when the vehicle is accelerating, not update the transformation matrix when the vehicle is decelerating, not update the transformation matrix when the vehicle is turning, and not update the transformation matrix when the vehicle is on a road surface that is not flat. 
   
     
     
         2 . The system of  claim 1 , further comprising an autonomous driving module configured to perform one or more autonomous driving operations based on the transformation matrix. 
     
     
         3 . The system of  claim 1 , wherein the alignment module is configured to determine whether at least one enablement condition is met prior to calculating the yaw, roll, and pitch of the LIDAR coordinate system. 
     
     
         4 . The system of  claim 1 , wherein the alignment module is configured to identify the lane markings based on an intensity threshold. 
     
     
         5 . The system of  claim 4 , wherein the alignment module is configured to dynamically determine the intensity threshold based on intensities of points of data in a predetermined region of interest. 
     
     
         6 . The system of  claim 1 , wherein the alignment module is configured to determine the lane marking direction using principal component analysis. 
     
     
         7 . The system of  claim 1 , wherein the alignment module is configured determine a normal to the ground plane and calculate the roll and the pitch based on the normal to the ground plane. 
     
     
         8 . The system of  claim 7 , wherein the alignment module is configured to determine the normal to the ground plane using principal component analysis. 
     
     
         9 . The system of  claim 1 , wherein the alignment module is configured to identify statistical outliers in the calculated yaw, roll, and pitch and remove the statistical outliers from the update to the transformation matrix. 
     
     
         10 . A method of performing LIDAR-to-vehicle alignment, the method comprising:
 collecting points of data provided based on outputs of one or more LIDAR sensors;   identifying lane markings based on the points of data;   determining a lane marking direction based on the identified lane markings;   calculating a yaw of a LIDAR coordinate system relative to a vehicle coordinate system based on the determined lane marking direction;   identifying a ground plane based on the points of data;   calculating a roll and pitch of the LIDAR coordinate system relative to the vehicle coordinate system based on the identified ground plane;   updating a transformation matrix based on the calculated yaw, roll, and pitch of the LIDAR coordinate system, and   not updating the transformation matrix when the vehicle is accelerating, not updating the transformation matrix when the vehicle is decelerating, not updating the transformation matrix when the vehicle is turning, and not updating the transformation matrix when the vehicle is on a road surface that is not flat.   
     
     
         11 . The method of  claim 10 , further comprising performing one or more autonomous driving operations based on the transformation matrix. 
     
     
         12 . The method of  claim 10 , further comprising determining whether at least one enablement condition is met prior to calculating the yaw, roll, and pitch of the LIDAR coordinate system. 
     
     
         13 . The method of  claim 10 , further comprising identifying the lane markings based on an intensity threshold. 
     
     
         14 . The method of  claim 13 , further comprising dynamically determining the intensity threshold based on intensities of points of data in a predetermined region of interest. 
     
     
         15 . The method of  claim 10 , further comprising determining the lane marking direction using principal component analysis. 
     
     
         16 . The method of  claim 10 , further comprising determining a normal to the ground plane and calculating the roll and the pitch based on the normal to the ground plane. 
     
     
         17 . The method of  claim 16 , further comprising determining the normal to the ground plane using principal component analysis. 
     
     
         18 . The method of  claim 10 , further comprising identifying statistical outliers in the calculated yaw, roll, and pitch and removing the statistical outliers from the update to the transformation matrix. 
     
     
         19 . A LIDAR-to-vehicle alignment system, comprising:
 a sensor data collection module configured to collect points of data provided based on outputs of one or more LIDAR sensors;   an alignment module configured to
 identify lane markings based on the points of data, 
 determine a lane marking direction based on the identified lane markings, 
 calculate a yaw of a LIDAR coordinate system relative to a vehicle coordinate system based on the determined lane marking direction, 
 identify a ground plane based on the points of data, 
 calculate a roll and pitch of the LIDAR coordinate system relative to the vehicle coordinate system based on the identified ground plane, 
 update a transformation matrix based on the calculated yaw, roll, and pitch of the LIDAR coordinate system, and 
 not update the transformation matrix when the vehicle is accelerating, not update the transformation matrix when the vehicle is decelerating, not update the transformation matrix when the vehicle is turning, and not update the transformation matrix when the vehicle is on a road surface that is not flat, 
   an autonomous driving module configured to perform one or more autonomous driving operations based on the transformation matrix,   wherein the alignment module is configured to determine whether at least one enablement condition is met prior to calculating the yaw, roll, and pitch of the LIDAR coordinate system,   wherein the alignment module is configured to identify the lane markings based on an intensity threshold,   wherein the alignment module is configured to determine the lane marking direction using principal component analysis,   wherein the alignment module is configured determine a normal to the ground plane and calculate the roll and the pitch based on the normal to the ground plane, and   wherein the alignment module is configured to identify statistical outliers in the calculated yaw, roll, and pitch and remove the statistical outliers from the update to the transformation matrix.   
     
     
         20 . The system of  claim 19 , wherein the alignment module is configured to dynamically determine the intensity threshold based on intensities of points of data in a predetermined region of interest, and
 wherein the alignment module is configured to determine the normal to the ground plane using principal component analysis.